Trim whitespace from PDF metadata

Guarantee that every PDF metadata field is trimmed of whitespace
characters, so users of the metadata do not have to worry about checking
themselves.

Consequently, change PDFiumEngine::GetMetadataByField() to
PDFiumEngine::GetTrimmedMetadataByField().

Bug: 93619
Change-Id: Ie51c7099d147e50e4e507e43fcca3a17fbba183a
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2609655
Commit-Queue: Daniel Hosseinian <dhoss@chromium.org>
Reviewed-by: K. Moon <kmoon@chromium.org>
Cr-Commit-Position: refs/heads/master@{#841712}

pdf/out_of_process_instance.cc

@@ -2374,7 +2374,7 @@ void OutOfProcessInstance::SendMetadata() {
   metadata_message.Set(pp::Var(kType), pp::Var(kJSMetadataType));
 
   const std::string& title = engine()->GetDocumentMetadata().title;
-  if (!base::TrimWhitespace(base::UTF8ToUTF16(title), base::TRIM_ALL).empty())
+  if (!title.empty())
     metadata_message.Set(pp::Var(kJSTitle), pp::Var(title));
 
   metadata_message.Set(

pdf/pdfium/pdfium_engine.cc

@@ -3950,19 +3950,22 @@ void PDFiumEngine::LoadDocumentMetadata() {
 
   // Document information dictionary entries
   doc_metadata_.version = GetDocumentVersion();
-  doc_metadata_.title = GetMetadataByField("Title");
-  doc_metadata_.author = GetMetadataByField("Author");
-  doc_metadata_.subject = GetMetadataByField("Subject");
-  doc_metadata_.creator = GetMetadataByField("Creator");
-  doc_metadata_.producer = GetMetadataByField("Producer");
+  doc_metadata_.title = GetTrimmedMetadataByField("Title");
+  doc_metadata_.author = GetTrimmedMetadataByField("Author");
+  doc_metadata_.subject = GetTrimmedMetadataByField("Subject");
+  doc_metadata_.creator = GetTrimmedMetadataByField("Creator");
+  doc_metadata_.producer = GetTrimmedMetadataByField("Producer");
 }
 
-std::string PDFiumEngine::GetMetadataByField(FPDF_BYTESTRING field) const {
+std::string PDFiumEngine::GetTrimmedMetadataByField(
+    FPDF_BYTESTRING field) const {
   DCHECK(doc());
 
-  return base::UTF16ToUTF8(CallPDFiumWideStringBufferApi(
+  base::string16 metadata = CallPDFiumWideStringBufferApi(
       base::BindRepeating(&FPDF_GetMetaText, doc(), field),
-      /*check_expected_size=*/false));
+      /*check_expected_size=*/false);
+
+  return base::UTF16ToUTF8(base::TrimWhitespace(metadata, base::TRIM_ALL));
 }
 
 PdfVersion PDFiumEngine::GetDocumentVersion() const {

pdf/pdfium/pdfium_engine.h

@@ -608,9 +608,9 @@ class PDFiumEngine : public PDFEngine,
   // document is loaded.
   void LoadDocumentMetadata();
 
-  // Retrieves the unparsed value of |field| in the document information
-  // dictionary.
-  std::string GetMetadataByField(FPDF_BYTESTRING field) const;
+  // Retrieves the value of |field| in the document information dictionary.
+  // Trims whitespace characters from the retrieved value.
+  std::string GetTrimmedMetadataByField(FPDF_BYTESTRING field) const;
 
   // Retrieves the version of the PDF (e.g. 1.4 or 2.0) as an enum.
   PdfVersion GetDocumentVersion() const;

pdf/test/data/document_info.in

@@ -18,13 +18,13 @@ endobj
 endobj
 {{object 4 0}} <<
   /Author (Chromium Authors)
-  /CreationDate (D:20200205153912+00'00')
+  /CreationDate (D:20200205153912+00'00' )
   /Creator (Your Preferred Text Editor)
   /Keywords (testing,chromium,pdfium,document,info)
   /ModDate (D:20200206094234+00'00')
   /Producer (fixup_pdf_template.py)
   /Subject (Testing)
-  /Title (Sample PDF Document Info)
+  /Title (   Sample PDF Document Info   )
 >>
 endobj
 {{xref}}